Laboratory Data Management - stat.smmu.edu. Linda Lab Data Management...DM Flow. 16 3.1 Central Lab Data Management

  • Published on
    14-Feb-2018

  • View
    213

  • Download
    1

Embed Size (px)

Transcript

  • The 2nd Clinical Data Management Training

    Laboratory Data Management

    September, 2010 at SMMU, Shanghai

  • 22

    Overview1

    Types of Laboratories2

    Lab Data Management3

    Standards4

    Agenda

    Key Messages5

  • 33

    1. Overview

    The vast majority of clinical studies use laboratory to provide:

    Safety Data Hematology, Chemistry, Urinalysis

    Efficacy Data Cholesterol level for Hyperlipidemia patient Plasma Glucose, for diabetes patient

    Special Data PK / PD data Genomic data Biomarkers

  • 44

    1. Overview

    There are many challenges in the data management task, but the ultimate challenge is managing the Lab Data.

    Laboratory data derived from tests of both drug safety and efficacy have been estimated to be as much as 60-80% of the data generated during the conduct of clinical trials Lab results tell the physician significant information about the body that are directly related to the safety and well-being of the subjectLab data are more difficult to interpret if one does not have clinical training, and what is appropriate to query is less clearIf involving central lab, data transfer and loading, it requires data manager having more understanding in database structure and data transfer technical and processes

  • 55

    Overview1

    Types of Laboratories2

    Lab Data Management3

    Standards4

    Agenda

    Key Messages5

  • 66

    2.1 Types of Laboratories

    Central LabCentral LabCentral LabA central lab processes samples from multiple clinical sites or studies at one central location. It often supports multicenter and international studies.

    Local LabLocal LabLocal LabA local lab is in close proximity to individual clinical study sites or patients and are most often used when timely results are needed.

    Virtual CentralLab

    Virtual CentralVirtual CentralLabLab

    A virtual central lab is typically a group of labs located throughout the world that are under the umbrella of one company. It is based on a central calibration that runs in parallel with samples from all labs in a clinical study.

  • 77

    2.1 Types of Laboratories

    Core LabsCore LabsCore LabsA core lab specialize in a particular therapeutic area or body system. Examples include stem cell core lab, ECG core lab, imaging core lab, cardiovascular core lab, etc.

    Specialty LabSpecialty LabSpecialty LabA specialty lab is used to analyze samples or run assays for non-traditional tests, which typically take a considerable amount of time and effort to produce. Examples include biomarkers, genetic testing, isolation of cancer genes, etc.

  • 88

    2.2 Pros and Cons of Different Labs

    Pros1. Uses one set of analytical

    equipment, methodologies, kits and reagents.

    2. Provides training and instructions for collection and shipping of samples.

    3. Standardized results from one set of reference ranges and units.

    4. Typically transfer data electronically from lab to the sponsor.

    Central lab

    Cons1. Very expensive due to

    logistical support and sample shipping.

    2. The turnaround time needed to receive central lab test data may be too long when immediate results are needed.

  • 99

    2.2 Pros and Cons of Different Labs

    Pros1. Lower costs and shorter

    turnaround time due to not having to ship samples.

    2. Local lab experience with processing samples from their subject population.

    3. Quick availability of test data, especially where the results could be the deciding factor on screening, dosing, etc.

    Local lab

    Cons1. Greater potential for errors

    due to paper-based data transcription.

    2. Differences between reference ranges from one lab to another.

    3. Variability in the methods used to perform tests.

    4. Reference ranges may be difficult to obtain.

    5. Time-consuming in verifying reference ranges.

  • 1010

    2.2 Pros and Cons of Different Labs

    Pros1. Reduced shipping costs.2. Simpler data processing

    due to having a central calibrator.

    3. Standardized results from one set of reference ranges and units.

    Virtual Central Lab

    Cons1. Requires detailed process

    and quality control measures to ensure lab results are reproducible with minimal variance from site to site.

  • 1111

    2.2 Pros and Cons of Different Labs

    Pros1. Highly experienced and

    qualified for performing specialty tests.

    Specialty Lab

    Cons1. Many specialty tests

    requires more time to generate test results.

  • 1212

    2.2 Pros and Cons of Different Labs

    Pros1. More focused quality

    control, more accurate results and a higher degree of standardization and specialization within a designated area.

    Core Lab

    Cons1. Additional time may be

    incurred for centralized processing.

  • 1313

    Central Lab Local Lab

    Sample Handling

    All samples shipped to the central lab for analysis and reporting

    Samples are analyzed at the local lab. No shipping

    Cost Expensive Relatively inexpensiveResults availability

    Need extra time due to sample shipping Quick

    Reference Range and Units

    One set of reference range and units

    Multiple lab-specific reference ranges and units

    Data handlingUsually data can be electronically transferred and uploaded into the database

    Data needs to be manually entered into the database

    Data Quality High quality of data Prone for transcription error

    Data Analysis Units conversion not needed and results are comparable Need units conversion

    2.3 Central Lab Vs. Local Lab

  • 1414

    Overview1

    Types of Laboratories2

    Lab Data Management3

    Standards4

    Agenda

    Key Messages5

  • 1515

    StudyStartUp Conduct Closeout

    ProtocolDevelopment

    CRFDevelopment

    DataManagement

    Plan

    DataExtraction

    DatabaseLock

    Dataanalysis

    ClinicalStudyReport

    DevelopDatabase

    DataKeyIn

    ExternalDataLoadingIn

    DataQualityReview

    MedicalReview

    Coding

    SAEReconciliation

    AnyQuery?

    QAstaffQualityControl

    DatabaseQualityControlReport

    Yes No

    DMsendQueryReport

    SiteRespondQueries

    UpdateDatabase

    DM Flow

  • 1616

    3.1 Central Lab Data Management

    Start Up PhaseProtocol Development

    What tests is required and when (Schedule of Assessment

    Central lab or local labCase Report Design

    Only a few questions are required Requisition Number (Accession No., Sample ID) Sample Collection Date Sample Collection Time Results are NOT needed to collect

  • 1717

    CRF Example _ Central Lab

  • 1818

    3.1 Central Lab Data Management

    Start Up PhaseDetermine if data should be loaded into the clinical databaseIdentify and involve lab vendor as early in the process as possibleIdentify key individuals for communication and follow throughCollect the reference range and units from the central lab

  • 1919

    3.1 Central Lab Data Management

    Start Up PhaseEstablish the procedures for collecting, transferring, loading, validating and editing external data and document.

    Data Transfer Agreement (DTA) File Format Specifications (FFS) Data Cleaning Plan

    Perform test transfer successfully before production

  • 2020

    3.1.1 Data Transfer Agreement (DTA)

    DTA defines:The format of files (Excel spreadsheet, ASCII, SAS dataset, text file, etc)Frequency of data transferFile naming conventionsEncryption and method of transfer (password encrypted email attachment, CD, secure FTP, etc)Type of transfer (accumulative vs. incremental)Primary and secondary contacts of sendersPrimary and secondary contacts of recipientsTransfer confirmation agreement

  • 2121

    3.1.2 File Format Specifications (FFS)

    FFS defines:File naming conventions

    File format Excel spreadsheet, ASCII, SAS dataset, text file, etc

    The delimiter (|, ;, || or blank)

    List of the variables and their order

    Type and length of each variable (char, numeric, date)

    Column position and field justification (for ASCII files) Key variables, which uniquely describe each sample

    record (study id, site/inv id, subjid, visit, sample id)

  • 2222

    3.1.2 File Format Specifications (FFS)

    FFS defines:Contents

    Test name, long and short (BUN vs. Urea)

    Units (conventional vs. SI units)

    Date and time format (YYYYMMMDD, DDMONYYYY, HH:MM 24hr)

    Handling of >,

  • 2323

    3.1.2 File Format Specifications (FFS)

    FFS defines:Unexpected/Unscheduled lab data

    Procedures for Database Updates New data

    Updates to already loaded data

    Procedures for ensuring blinding

  • 2424

    3.1.3 Data Cleaning Plan

    Data checks that can be applied include:Reconciliation between CRF data and loaded lab data

    Demographic (subj. ID, subj. initials, date of birth, sex) Sample collection (visit name, collection date and time,

    sample id)

    Missing individual test result within a panel Missing RBC result in Hematology test at screening visit

    Missing / invalid reference range and unitsOut-of-range values (against medical history or adverse event)Inclusion/exclusion criteria involving lab dataDuplication test results

  • 2525

    3.1 Central Lab Data Management

    Conduct PhasePerform production transfers as per DTAResolve loading problems with the lab vendor, if anyReview and clean the external lab data on an ongoing basis according to data cleaning planResolve data discrepancies with the site and the lab vendorMaintain the reference range and units and perform change control

  • 2626

    3.1 Central Lab Data Management

    Close-down PhasePerform the last production transfer which contains all results of all lab samplesResolve loading problems with the lab vendor, if anyReview and clean all the external lab dataResolve all data discrepanciesRelease blinded data, if appropriateFile storage and archiving

  • 2727

    3.2 Local Lab Data Management

    Start Up PhaseProtocol Development

    What tests is required and when (Schedule of Assessment

    Central lab or local labCase Report Design

    Requisition Number (Accession No., Sample ID) Sample Collection Date Sample Collection Time Results are collected on CRF

  • 2828

    CRF Example _ Local Lab

  • 2929

    3.2 Local Lab Data Management

    Start Up PhaseCollect the reference range and units from the individual local labsVerify the reference ranges and unitsDTA and FFS are not needed but data cleaning plan is still required

  • 3030

    Data checks that can be applied include:Reconciliation between CRF data and loaded lab data NOT neededMissing individual test result within a panel

    Missing RBC result in Hematology test at screening visit

    Missing / invalid reference range and unitsOut-of-range values (against medical history or adverse event)Inclusion/exclusion criteria involving lab dataDuplication test results

  • 3131

    3.2 Local Lab Data Management

    Conduct PhaseResolve data discrepancies with the siteMaintain the reference range and units and perform change control

    Close-down PhaseReview and clean all the lab dataResolve all data discrepancies

  • 3232

    Overview1

    Types of Laboratories2

    Lab Data Management3

    Standards4

    Agenda

    Key Messages5

  • 3333

    4. Standards in Lab Data Collection & Interchange

    StandardsStandards

    Test Name & Units

    CRF design

    Data Structure

    Data Interchange

  • 3434

    4.1 Test Name and Units

    Test NameWBC vs. White Blood Cell; BUN vs. Urea; AST vs. Aspartate transaminase vs. SGOTStandards

    CDISC standard - Terminology ( ) LOINC Codes () SNOMED CT ()

    CDISC Clinical Data Interchange Standards Consortium

    LOINC Logical Observation Identifies Names and Codes

    SNOMED CT Systematized Nomenclature of Medicine Clinical Terms

  • 3535

    CDISC PT(LBTESTCD)

    Long Name /Descripti

    on (LBTEST)

    Definition Synonyms

    BILI BilirubinA measurement of the total

    bilirubin in a biological specimen.

    Total Bilirubin

    BILIND Indirect Bilirubin

    A measurement of the unconjugated or non-

    water-soluble bilirubin in a biological specimen.

    BUN Blood Urea Nitrogen

    A measurement of the urea nitrogen in a blood

    specimen.

    CDISC Terminology (Laboratory Data) - example

  • 3636

    LOINC Codes - example

    LOINCCode

    Component

    Property

    Time Aspect

    System Scale Method LOINCShort name

    14631-6 BILIRUBIN

    SCNC PT SER/PLAS

    QN BilirubSerPl-

    sCnc1974-5 BILIRU

    BINMCNC PT FLU QN Bilirub

    Fld-mCnc

    1977-8 BILIRUBIN

    ACNC PT UR ORD BilirubUr Ql

    Component analyte being measured; Property observed Time Aspect time of measurement; System specimenScale Quantitative, qualitative, ordinal; Method where applicable

  • 3737

    4.1 Test Name and Units

    Test UnitsReported UnitsConventional Units (typically based on US measuring methods)SI Units (le Systeme International dUnites)

    For example, reference and units for Potassium (

    3.5 - 5.0 mmol/L 3.5 - 5.0 mEq/L 13.7 - 19.5 mg/dL

  • 3838

    4.2 CRF Design Standards (CDISC CDASH)

    Central processingLab status (whether or not lab sample was collected)Date of collectionTime of collectionPanel name (e.g. Chemistry, Hematology, Urinalysis)Planned time point Protocol-defined testing conditions met (e.g Fasting)Accession number

    Only the highly-recommended variables are listed

    CDISC Clinical Data Interchange Standards Consortium

    CDASH Clinical Data Acquisition Standards Harmonisation

  • 3939

    4.2 CRF Design Standards (CDISC CDASH)

    Local processingLab status (whether or not lab sample was collected)Date of collectionTime of collectionPanel name (e.g. Chemistry, Hematology, Urinalysis)Planned time point Protocol-defined testing conditions met (e.g Fasting)Test nameTest resultUnits

    Only the highly-recommended variables are listed

  • 4040

    4.3 Data Structure Standards (CDISC SDTM)

    LB domain

    One record per lab test per time point per visit per subject

    See the example in the next two slides

    For more details, please visit CDISCwebsite: http://www.cdisc.org/sdtm

    SDTM Study Data Tabulation Model

    http://www.cdisc.org/sdtm

  • 4141

    Variable Name

    Variable Label

    Type

    Controlled Terms, Codelist or Format

    Role CDISC Notes Core

    References

    STUDYIDStudy Identifier

    Char IdentifierUnique identifier for a study. Req

    SDTMIG2.4.4

    DOMAIN

    Domain Abbreviation

    Char LB Identifier

    Two-character abbreviation for the domain.

    Req

    SDTMIG2.4.4, SDTMIG4.1.2.2, SDTMIGAppendixC2

    USUBJID

    Unique Subject Identifier

    Char Identifier

    Identifier used to uniquely identify a subject across all studies for all applications or submissions involving the product.

    Req

    SDTMIG2.4.4, SDTMIG4.1.2.3

  • 4242

    Variable Name

    Variable Label Type

    Controlled Terms, Codelistor Format

    Role CDISC Notes Core References

    LBSEQSequence Number

    Num Identifier

    Sequence Number given to ensure uniqueness of subject records within a domain. May be any valid number.

    Req SDTMIG2.4.4

    LBGRPID Group ID CharIdentifier

    Used to tie together a block of related records in a single domain for a subject.

    Perm

    SDTMIG2.4.4, SDTMIG4.1.2.6

    LBORRES

    Result or Finding in Original Units

    Char

    Result Qualifier

    Result of the measurement or finding as originally received or collected.

    Exp

    SDTMIG2.4.3, SDTMIG4.1.5.1

  • 4343

    4.4 Data Interchange Standards

    It is estimated that the cost to the industry per year simply for laboratory data interchange itself is at least $150m and that between approximately 30% and 60% of that cost could be saved from the use of a standard

    The existing standard models for the interchange of laboratory data includes:

    ASTM InternationalHealth Level 7 (HL7)ACDM (the Associate for Clinical Data Management)X12

  • 4444

    4.4 Data Interchange Standards

    CDISC standard Laboratory Data ModelThe default implementation of the LAB Model is bar delimited ASCII The superset of data fields are in 12 levels;

    Good Transmission Practice Study Site Investigator Subject Visit Accession Record Extension Type Base Specimen Base Battery Base Test Base Result

  • 4545

    Example of lab data file

  • 4646

    4.4 Data Interchange Standards

    CDISC standard ODMA vendor neutral, pla...

Recommended

View more >